Majalah Ilmiah UNIKOM

Vol.8, No. 1

10

H a l a m a n

Muhammad Aria

Ulysses16 cases, AG takes 6.47 times

slower than SA and 1.3 times than GA

algorithm. As the comparisons of the

memory complexity, interval type-2 TSK

Fuzzy algorithm needs 13.88k bytes units of

memory, while SA only needs 10.67k bytes

unit memory but GA algorithm needs 31.11k

bytes units memory. This comparison result

is shown in Table 5.

CONCLUSION

This study began as a search for new ar-

chitecture for hybrid optimization based on

Simulated Annealing and Genetic Algorithm

called Annealing-Genetic Algorithm (AG). AG

can solve the TSP for large-scale problem.

We found that AG is much faster than Simu-

lated Annealing and that solution quality

(total elapsed time) is slightly better. Simula-

tion results showed that AG improves the

(a)

(b)

(c)

Figure 7. Example TSP Problem: TSP225

(a) Distribution City from TSP225

(b) Known Optimal Solution

(c) Best Solution found by us